Do vegans kill more animals through crop deaths?

The objection

“Harvesting crops kills animals. Combines crush mice and ground-nesting birds; ploughing destroys habitat; pesticides kill countless small creatures. A vegan eating field-grown grain isn’t bloodless, and a grass-fed cow might cause fewer deaths per calorie than a field of wheat.”

The answer

Crop farming does kill animals. Tractors crush field mice, combines destroy birds’ nests, fertiliser runoff poisons fish, and in Australia farmers lay poison to control mouse plagues [1]. The objection goes wrong in assuming that eating meat spares those animals, when much of the world’s harvest goes to farm animals: about 41% of the world’s cereals are fed to livestock, against 48% eaten by people (Our World in Data, 2021) [4]. Cut out animal products and the need for cropland falls, and Poore and Nemecek’s data from about 38,700 farms in 119 countries (2018) show that a world eating none would need 19% less arable land, and 76% less farmland in all [3].

How many animals die in crop farming?

Nobody knows with any confidence how many animals crop farming kills, and the figures usually quoted rest on very little. The best known, Steven Davis’s (2003) estimate of 15 animals killed per hectare a year, averaged two small studies, one of wood mice at a grain harvest and one of rats in Hawaiian sugarcane [1]. When the philosophers Bob Fischer and Andy Lamey (2018) reviewed the evidence, they found only one more study that recorded a field animal killed directly by harvesting [1]. Most of the radio-tracked wood mice that died after that harvest were taken by tawny owls, weasels and other predators once the crop cover had gone, and Davis treated the sugarcane deaths as a yearly rate although sugarcane takes two years to grow [1]. Michael Archer’s (2011) Australian estimate, at least 100 mice poisoned per hectare of grain a year, assumed a mouse plague on every hectare every four years, when plagues reach about 2.3% of the country’s grain land in an average year [1].

Fischer and Lamey’s own tentative figure for the US is about one animal killed per hectare of harvested cropland a year, leaving out the animals that predators take after harvest. Counting those would raise it a good deal, and counting insects and other small invertebrates would raise it enormously [1]. Their review ends with trends in farming that cause little or no harm to field animals, which they think could shrink the problem or end it [1].

Does eating meat cause fewer crop deaths?

Eating meat almost always means more crop deaths, because the animals eat crops too and give back only a fraction of them as food. In US data, pigs and poultry, fed on concentrates such as maize, returned about 9% and 13% of the calories in their feed (Shepon et al., 2016) [2]. Feed crops are sown and harvested like any other, so a meat-eater shares the field deaths behind the feed and adds the death of the animal itself. Fischer and Lamey (2018) note that the argument stands a better chance when it defends eating animals that have not been fed grain or other crops, since it then need not count the field animals killed to grow their feed [1].

Is grass-fed beef the exception?

Grass-fed beef is the argument’s best case, and even there the numbers lean against it, because beef needs so much more land than plant food for the same protein. Per 100 g of protein, beef from beef herds uses 164 square metres of land for a year on average, about 48 times the 3.4 for peas, and only a tenth of its production comes in under 42 (Poore and Nemecek, 2018) [3]. A death rate per hectare means little until it is multiplied by that land, so a pasture would have to kill a small fraction of what a pea field does just to break even. Grazing alone also feeds few people, since grazing-only systems supplied about 1 g of protein per person per day worldwide in 2000, against the 27 g a day that all land animals now supply, according to an Oxford-led review (Garnett et al., 2017) [5].

Nor is pasture free of killing. Improved pastures are sometimes mown for winter silage [5], and Fischer and Lamey cite a 1990 study of bobolinks, a small North American blackbird, killed when hay fields were mown [1]. The poison laid during Australia’s mouse plagues, which Archer counted against grain, protects pasture and stored hay as well [1].